Individuals working for a PE firm, collectively referred to as GPs, obtain revenues from several sources: i) management fees charged to LPs, ii) carried interest (retained from asset sale proceeds), iii) fees charged to the asset owned by the fund they advise, iv) expense reimbursement from the fund they advise, v) expense reimbursement from the assets owned by the fund they advise, vi) other related party transactions (kick-backs, discounts,…).
What do we know?
We can see PE fees in the annual reports of some LPs, sometimes going back a long time. The amount has not changed much over time and is similar across LPs: between 1% and 1.5% of NAV. This amount being of a similar order of magnitude as fees charged by active public equity managers, it never raised a red flag. However, this low amount is due to the fact that LPs report only the fees directly charged to them (item i listed above). In addition, LPs are not charged the full amount of management they owe. GPs reduce the amount of management fees they charge as a function of how much they charge to the asset (item iii) and may waive management fees in exchange of a higher carried interest (item ii). Moreover, LPs who invested in PE via a fund-of-funds may have reported only the fees charged by the fund-of-funds and not those of the funds that the fund-of-funds invested into. The logic is the same: LPs simply reported what the total they were invoiced for in a given year.
Academics simulated fee amounts paid by LPs using different approaches, but restricted themselves to the first three items above. The problem is that many assumptions need to be made. Even knowing the LPA (the contract between the LP and GP) is not sufficient to have a precise idea because the amount for item iii is not specified in the LPA; it is at the discretion of the GP. In addition, the exact timing of asset divestment and returns obtained will determine how much is charged for item i and ii. As a result, we do not know the exact amount represented by the first three items. A partial exception is that we can see these amounts in the annual reports of the publicly listed GPs (e.g. Blackstone) but they have been listed only for ten years, only a handful of them are, and some assumptions are still needed because the information provided is not always sufficient.
A rough estimate seems quite close to all the bits and pieces of information collected from all of the above. Capital committed: $100. A management fee of 2% is paid over five years: $10, then on average only half the amount is invested over a five year period leading to a fee of $5. $6 is charged to the asset (item iii) with 80% rebated against management fees. Hence, say $10 of management fee is charged, plus $6 from item iii. Fund doubles money gross of fees leading to a carried interest of (100-10)*20% = $18. Total is $34. (100-16)=84 was invested for 4.5 years (used to be four, and recently closer to five on average) and grew to (200-18)=182, hence average NAV was $133. The expense ratio is 34/133/4.5= 5.7%. LP gave $110 and received $182 (multiple of 1.65, rate of return of 12%). If LPs report only the fees they have been charged for in this example, they would report 10/133/4.5= 1.6% p.a. This is rough, but again all these statistics are extremely close to everything we have observed so far (including the returns). In this simple example, LPs earned something close to the average listed stock return (see previous blog) and paid 5.7% p.a. of fees.
Only the first three items are covered in this estimate. Item iv (fund expenses) is not negligible; it includes organizational expenses (cost of setting up the fund), the cost of subscription lines (credit lines used by funds to bridge finance between investments and capital calls), etc. We do not know anything about this item. We cannot even simulate it. The amount is given in the Fund quarterly reports (if at all). A fund I recently reviewed charged $2 of fund expenses for $100 committed, which would raise the bill in the above example to $36 and the expense ratio to 6% p.a.
Items v and vi are yet more difficult to trace and quantify. Yet, the difficulty on this front is deeper than the mere accounting of these expenses. The core of the issue lies in the difficulty in drawing a line between what is an operational expense and a fee. Take the example of an activist hedge fund that forces a listed company to hire an expensive consultant to fix operations. It would seem odd to add the cost of the consultant to the expense ratio of the activist hedge fund. Yet, if a GP uses internal teams to fix operations, then this could count towards its expense ratio. If I had to boil it all down, the very core issue is with the notion of arm’s length prices. Determining an arm’s length price is very difficult and at the centre of many pressing problems, such as international taxation issues. If GPs were deriving no utility from any of the services they render to the asset they manage, then one could argue that an appropriate expense ratio is simply item i after any rebate plus item ii. In this case, fees would be much lower. In the working example above, they would be equal to: 28/133/4.5= 4.7%. If GPs derive a utility (e.g. a profit) from these expenses and other related party transactions, then the excess should be counted as a fee. Making this estimation is therefore crucial to know the exact fee bill, but extremely difficult.
Why do we care?
There are two reasons why one may care. The first reason has to do with fairness. Is it fair that a pension fund gave $10 billion, received $16.5 billion back 4.5 years later, earning a 12% return, and paid out more than $3.6 billion of fees during that period?
The second reason is that fees are certain while returns are not. If future returns of PE were lower for some reason (e.g., increased competition, new entrants, overall lower expected returns) then knowing how much fees are and how they are split between fixed and return-dependent is key. Let’s reduce the amount generated from $200 to $150 in our working example. Carried interest is reduced to (50-10)*20%=$8. Other fees are the same, but average NAV was lower. Restricting the calculations to item i to iii, we obtain a total of $24, i.e. 4.7% p.a. for a return of 6% p.a. Now the $10 billion investment brought back $13 billion, i.e. 6% p.a. and $2.4 billion in fees were paid. Obviously, if returns turn negative then the situation looks worse with a fee bill of $1.6 billion for a fund that lost money.
We care because we do not live in past returns but need to be forward looking. In addition, ethics and fairness might be a concern to some.